Accountable Care Organizations: Prospects in the New Era of Health Care Reform

An Accountable Care Organization (ACO) is a relatively new patient care model that has received much attention over the past few months, particularly in the context of recent health care reform efforts attempting to "bend the cost curve" of growing national health expenditures. First coined in late 2006 by Fisher et al, an ACO may be described as a group of geographically linked providers that are accountable for the cost and quality of care for a specific patient population within a region. Under this model, providers such as primary care physicians, specialists, and hospitals would coordinate care in a longitudinal fashion, thereby improving quality and producing tangible cost-savings. These providers would then be reimbursed for a certain percentage of their total savings, or be held accountable for their inefficiencies and waste. Recently the ACO definition has been broadened to include “virtual networks” of providers, as opposed to distinct, extant organizations. These loosely connected regional networks could improve patient care through use of electronic health records (EHRs), better coordination, e-prescribing, fewer re-hospitalizations, and a stronger emphasis on primary care, allowing them to rein in rising costs, and then share in resulting savings as an incentive. Essentially, the ACO framework is modeled off the successes of integrated delivery systems like Geisinger and Intermountain Health Care, and may now be extended to virtual networks that only cover a specific subset of the patient population.

The Patient Protection and Affordable Care Act (ACA), signed into law by President Obama this past March, defines an ACO as any network of physicians or individual practices, physician-hospital organizations (PHOs), or hospitals that meets specific requirements as set forth by Section 3022 and the Secretary of Health and Human Services (HHS). The Secretary will work with networks “willing to become accountable for the quality, cost, and overall care of the Medicare fee-for-service beneficiaries assigned to it” for at least a three year period. Medicare ACOs would be responsible for most Part A and B items and services, and would be obligated to incorporate a sufficient number of primary care physicians to care for a minimum of 5,000 Medicare beneficiaries. ACOs must also be able to implement and report specific quality, performance, and cost measures in order to be eligible for this “shared savings program.” If an ACO can reduce total spending in relation to average Medicare expenditures per capita (after adjustment for beneficiary characteristics), then that organization will be able to share the savings. In other words, if an ACO meets all the applicable requirements, the Secretary will define a certain percentage of the cost-savings that may be reimbursed, using a partial capitation or related payment system. However, reimbursements will not be doled out if the formation of an ACO results in greater spending than would otherwise be expended if the model were not implemented.

Section 3022 of ACA, which encourages the formation of ACOs under the Medicare program, is only five pages in length. This may be attributed to the novelty of the ACO concept. Therefore, framework suggestions and implementation strategies must be found in the literature, which is sparse. Most agree that coordinating care and better oversight among providers will improve quality of care for patients, and there is empirical data to suggest that organized multispecialty medical groups with this goal in mind have achieved just that. However, in order to implement this longitudinal model of care, many potential ACOs will require significant technical assistance. All but the most advanced and integrated providers will require the establishment of “organizational, legal, financial, and budgeting relationships with payers” in order to create the overarching ACO framework, as well as mechanisms specific to meeting performance and quality standards. Stakeholders would require significant practice redesign, focused on leadership development, EHR implementation, and quality improvement. Since this type of organizational structure cannot be created overnight, researchers have suggested different means of implementation.

One such strategy demonstrates how the ACO model could be achieved in stages, slowly increasing organization, performance, and risk. One of the difficulties of developing a new ACO is the acquisition of associated downside risk. In this implementation approach, a Level 1 ACO would receive reimbursement if shared savings were achieved, but would not be responsible for the fallout if these initial conditions were not met. As a more advanced infrastructure is created, the ACO would become accountable for an increasing share of the costs if targets were not met, eventually becoming fully responsible. Other variations on this approach have been suggested, but again the major impediment confronting a fledgling ACO is the initial acquisition of risk. There is little doubt that better coordination of care can improve quality, but providers are unlikely to invest in this framework without proof of financial success or added financial backing. A number of pilot programs in states such as Vermont and Massachusetts have begun with these goals and limitations in mind.

ACA, however, mentions nothing about gradual implementation, leaving the overall execution plan vague. The statute states that that a contracted ACO will be “at financial risk for some, but not all, of the items and services covered under parts A and B.” This implies that ACOs which contract with the Secretary of HHS are likely to have already jumped many of the hurdles impeding implementation. These groups are probably already integrated health systems, or ones likely to become integrated in the very near future.

For this reason, some researchers have postulated that Academic Health Centers (AHCs) are in an excellent position to lead the ACO charge. Some AHCs arguably have a unique ability to become ACOs due to the broad social, political, and economic power that many hold, and the access to resources that independent physician practices and smaller PHOs may not have. In order for AHCs to become leaders in this arena, however, they would have to fully commit to primary care within the community, potentially through the development of patient-centered medical homes. Researchers suggest that fragmentation, emphasis on research, and higher costs due to medical education may hinder the progress of AHCs in reaching this goal. Thus, incentivizing AHCs to become leaders in this field most likely will follow upon successes of current ACO projects and demonstrations, and will subsequently require federal subsidies and a great deal of funding for broader integration and implementation.

Critics of ACOs point to the lack of empirical data for their structure. Some note that the reward of “partial capitation,” while perhaps efficacious in reducing spending, lacks evidence for significant improvement in the quality of patient care. Even though performance measures are included in ACA, ACOs will have to prove that these two concepts are linked. Dove et al also note that health information technology (HIT) and EHRs will need to deliver on their lofty expectations to improve quality of care, otherwise ACOs will find themselves holding expensive technology yielding few benefits. Others remain skeptical of the “virtual” framework, noting that though this model might be theoretically aesthetic, it lacks pragmatism. Since an ACO is fundamentally defined by geography, integration and coordination of care will require cooperation among competing physicians and their practices, a model of physician collegiality that some critics deem unlikely. While existing integrated health systems might see some success, the rest of the health care delivery system may not follow suit. Still others question the role of AHCs in the ACO framework. Qualitative evidence shows that large integrated provider groups, such as AHCs, dominate negotiations with private insurers, causing premiums and payment rates to soar to levels much higher than Medicare. In other words, by squeezing one end of the balloon, the other end expands, inevitably hurting consumers and private insurers. This raises several antitrust issues, which, despite legal precedents favoring the ACO concept, will require major investments in legal expertise by potential ACOs.

There is also the question of how consumers would be “assigned” to an ACO. Sixty percent of physicians doing only inpatient work are affiliated with a single hospital, and of the remaining 40%, three-quarters of their work is at a one “primary” hospital. Consequently, for physicians engaged in any inpatient work, 90 percent or more is at a “primary” hospital. A loose network could be built around this, and patients could be assigned accordingly. Questions remain, however, about whether patients should be informed of this “assignment.” In the case of a fully integrated PHO, patients generally have the option of leaving the network (as patients have the option of not joining an HMO or any other provider network), whereas when a loose network spans an entire region – as many of the ACO designers have in mind – patients may not be given a choice. Will they be subjected to this model without a means of opting out? Or do consumers even need to be informed? These pragmatic and ethical questions linger.

Despite being such a small part of this year’s health reform law, the ACO model has broad implications for variations in regional health spending, a controversial topic put forth by the Dartmouth Atlas of Health Care (“the Atlas”). Tracking the Care of Patients with Severe Chronic Illness was published by the Dartmouth Institute for Health Policy and Clinical Practice, arguing that in the United States, “more care does not equal better care” and “some chronically ill and dying Americans receive too much care: more than they or their families actually benefit from.” This study focused on Medicare beneficiaries who had at least one chronic illness, summing their total medical expenses in the last two years of life. Researchers then mapped these findings geographically, analyzing Medicare Part A and B spending per state, per hospital referral region (HRR), and for the 2,826 U.S. hospitals with at least 400 patient deaths between 2001 and 2005. Extrapolating from these results, many have argued that physicians in areas such as the upper Midwest offer cheaper, and thus better, care for their patients, as opposed to their colleagues in the South, where end-of-life care costs are much higher. “Too much care” correlated to increases of up to 52% in Medicare reimbursements between the highest and lowest spending regions. Further data suggests that “after adjustment for demographic and baseline health characteristics and changes in health status, the difference in [spending] between the highest and lowest quintiles can be reduced to 33%,” but the fact remains that at least a third of all expenditures in certain regions of the country may be deemed wasteful.

Fully integrated provider systems, such as the Geisinger Health System or the Group Health Cooperative of Puget Sound, may see moderate cost-savings results utilizing the ACO framework, but regions identified by the Atlas as high spending districts arguably could see the greatest reduction in expenditures if they buy into this model. The difficulty is, however, incentivizing these providers to form ACOs. Decreasing health care expenditures in these areas correlates with a decrease in profits for their physicians and hospitals. Even with partial reimbursement for substantial cost-savings, the question remains, will it be enough to improve their bottom-line? Should providers in these regions receive a higher initial percentage of reimbursement to draw them into the system, or does this simply reward years of waste? Should they begin as Tier 1 ACOs? Organizations that want to be held accountable for their care are not the ones the Secretary of HHS needs to court for ACO Medicare contracts; it is the providers that do not.

Furthermore, these studies acknowledge that there are large differences among regions of the country. Demographics, health characteristics, cost of living, and health outcomes, for example, vary throughout the United States. So will a “one-size-fits-all” model work nationally? If there are to be different agreements among providers and Medicare, how should these statistical variations be adjusted, or should they be adjusted at all? One of the attractive aspects of the ACO concept is the fluidity and malleability of the framework across provider groups of different sizes and levels of integration. Whether the Medicare payment process will achieve this same type of flexibility is yet to be determined.

Ultimately, ACA does not provide significant funding incentives for the formation of ACOs. There is funding in the health reform law and the American Recovery and Reinvestment Act of 2009 for EHRs and HIT, but subsidies to develop infrastructure for legal and leadership needs, as well as practice redesign are limited. Congress has to provide a larger carrot in the future if the ACO model is to become a national success, and pilots and demonstrations projects will have to achieve significant cost-savings in order for the model to catch on. And as far as shared savings, “the fundamental question is whether there are enough physicians interested in doing the hard work of forming and managing ACOs capable of directly managing even partial capitation risk. It is possible that for all of the theoretical advantages of this approach, there would be few takers in a voluntary ACO program using any form of capitation.” However, the future is not entirely bleak. Four years after the ACO model was first proposed, Congress has recognized its value and has provided legislative support for implementation of Medicare ACOs in an effort to decrease health care costs while concordantly improving quality. As this model evolves over time, look for innovations and potential successes; however, the status quo will remain until there is true payment reform and a larger commitment to expanding primary care.

(sources available upon request)